Re: emacs-29 1ef8b90ae06: Simplify imenu setup for {cmake, dockerfile}-ts-modes

[Po Lu]

Stefan Kangas <stefankangas@gmail.com> writes:

> ? They are not affected by this change though?  Are we looking at the
> same patch (commit 1ef8b90a)?
>
> Which areas of Emacs do you think are affected?  AFAICT, nothing risks
> breaking outside of the changed modes.  For example, autoloading is not
> affected (`declare-function' only affects byte-compilation).

You are correct as to autoloads, but it certainly affects imenu.

> Did you read Bug#68706?  Yuan decided to install it as a bugfix on
> emacs-29.
>
> The question is if this change is different from the many other bug
> fixes that we routinely install without much justification.  This is why
> I'm asking for something more specific.

Yes I did: it's larger, and encompasses much more than only the fix.
Yuan's precise wording was:

> Thanks Randy, pushed to emacs-29 since it includes a fix.

which does not suggest that the safety of the patch in whole was a
factor in his decision, but only the fact that a bugfix was one
component of the patch.  This prompted my reply that the inclusion of a
bugfix is not the sole criterion we expect changes to the release branch
to satisfy, and I trust we are in agreement in this respect.

> I don't know which changes this is in reference to, but I know that you
> often work on things like X support, that carries with it a bigger risk
> for breakage (as it's more complex), and it affects more users to boot.
> It is natural to be more careful with changes in core functionality than
> in specific modes.
>
> Note also that we have been decided to be more lax when it comes to the
> treesitter stuff, seeing as its new in Emacs 29.

No disagreements there, but "the treesitter stuff" doesn't extend to
imenu setup code in major modes which happen to make use of tree-sitter.